VERN™ is the Virtual Emotion Resource Network
We set out over a decade ago to opertaionalize machine capable emotion detection, understanding that emotional intelligence would be necessary for “human like” interactions. We’d all grown up fans of science fiction, and of particular interest to us were popular fictional computerized assistants like Iron Man’s “J.A.R.V.I.S.,” Batman’s “Bat-computer/Dupin“, and Star Wars’ C3PO and R2D2. These computerized companions understood their heroes’ emotions and were relatable because of an appearance of a “personality.”
Sometimes what lurks in fantasy belies an unmet demand. And in order to get there meant we would undertake a journey of research & development to create VERN™ AI. We’ve put years of software development, content analysis, mental health and medical experience to develop a system that is explainable, flexible, as non-biased as it can be, and accurately scores emotions in a way that people can understand. VERN™ AI isn’t built on yesterday’s news, it’s the emotion recognition software of tomorrow.
The Technology
Overview
VERN™ is developed to detect human emotions in communications. It predicts the user’s intent on a scale of 0-100%. This scale represents intensity of the emotion present. VERN™ is a patented system that detects latent emotional clues at the lexical level, and corpus level.
You can connect and use the API service or find VERN™ in your favorite AI Marketplace.
API
VERN™ provides a REST API to perform live analysis of a body of text. The results are returned in a JSON style format. An active API key is required to make a request and is available in the Dashboard.
On-Prem
VERN™ is available via marketplaces, and can be directly integrated into your ML Ops project or used as a stand-alone on your server. The packages vary depending on marketplace. For more details, please see our On-Premises solutions page for the latest in marketplace offerings.
The Concept
Overview
VERN™ detects distinct emotional signals from communications at the lexical level. We conceptualize emotions as being a mix of Euphoria, Dysphoria, and Fear. Our emotions model prediction error indicators and are described in brief below.
Anger
VERN™ provides analysis of Anger. This is conceptualized as communicating the recognition of a malignant incongruity. In layman’s terms, it would be recognition that something was not “fair” or would harm the sender of the message in some way.
Sadness
VERN™ provides analysis of sadness. This is conceptualized as the acceptance of a malignant incongruity. In layman’s terms, it’s the communication of dread and resignation.
Love & Affection
VERN™ provides analysis of Love & Affection. Unlike other Dysphoric emotions Love & Affection is the absence of an incongruity. In layman’s terms, this means that the sender’s goals align and they’re communicating this. It could be considered joy, grace, and virtuous.
Fear
VERN™ provides analysis of Fear signals. Fear is an interesting neuroscientific phenomena, that we’re excited about (really). While there is no current method to actually measure fear, VERN™ measures the sender’s fear response. Individuals communicate their fear, and that’s how all systems detect fear signals.
Humor
VERN™ provides analysis Humor. It’s conceptualized as the detection of a benign incongruity. In layman’s terms, it’s whatever doesn’t kill you makes you laugh.” The model incorporates the Incongruity Theory of Humor; with the Relief and Superiority Theories being a different types of incongruity. Currently, the team is undergoing experimentation with Humor so the analysis won’t be provided in this version of VERN™ AI. We will have more to come soon!
The Application
Human Emotion Detection
VERN™ provides analysis of Anger, Sadness, Fear, Love & Affection and Humor. We have found that VERN™ is successful in chatbots and virtual assistants, as a method of analysis for mental health applications, in analyzing internal and external communications (including human resources, marketing, social media and public relations).
VERN™ can be used in many ways and some of these can be found on the Use Case section of this website.
We can help!
If you would like to learn how VERN™ staff can assist in planning and executing your chatbot, virtual assistant and can recommend providers of voice to text and other AI/ML, or NLP/NLU tools to complete your VERN™ powered application. Click here to learn more!
Dedicated staff of professionals
Craig Tucker
Craig Tucker is the creator of VERN™ and a co-founder.
Bryan Novak
Bryan is the CTO of VERN™ and a co-founder.
Edward Christensen
Edward is the Counsel for VERN, LLC and is also a co-founder.
Brendan Watson
Dr. Watson is head of research for VERN™
Dennis Walters, II
Dennis is VERN™'s development guru & chief of its technical operation.
Sam Jebeli-Javan
Bachelor’s degree in Computer Science and a minor in International Management, with 23 years of Fintech experience at IBM, American Express and Citibank, with 6 years in Conversational AI. Led the business development and deployment of the first Banking chatbot at scale in 2019 with Citibank